Basic Appl. Ecol. 4, 249–262 (2003) © Urban & Fischer Verlag http://www.urbanfischer.de/journals/baecol

Basic and Applied Ecology

Patterns and processes of biological invasion: The Chinese mitten crab in San Francisco Bay Deborah A. Rudnick1*, Kathryn Hieb2, Karen F. Grimmer3, Vincent H. Resh1 1

University of California, Berkeley California Department of Fish and Game 3 Monterey Bay National Marine Sanctuary 2

Received January 23, 2002 · Accepted May 28, 2002

Abstract The Chinese mitten crab, native to coastal rivers and estuaries of central Asia, has invaded several European countries over the past century, causing widespread concern because of its periodically extreme abundance and burrowing behavior that causes bank erosion. San Francisco Bay and its tributaries contain the first and, currently, only known established population of the Chinese mitten crab in North America. Discovered in South San Francisco Bay in 1992, the mitten crab has spread rapidly to cover several thousand km2 surrounding the Bay. Between 1995 and 2001, we monitored distribution, population dynamics, and life history attributes of the Chinese mitten crab in San Francisco Bay and its tributaries. Burrow densities increased from a mean of 6 burrows/m2 in 1995 to > 30 burrows/m2 in 1999 in tidal portions of South Bay tributaries. Mitten crabs are associated with: tidally influenced portions of Bay tributaries as young juveniles; with freshwater streams (≤250 km from their confluence with the Bay) as older, migrating juveniles; and with the open waters of the Bay as reproductive adults after migrating from fresh water to reproduce between late fall and early spring. Population size peaked in 1998, with 750,000 crabs counted in fall migration in a North Bay tributary and 2.5 crabs/tow collected from North Bay breeding grounds; abundance subsequently declined greatly (2500 crabs in the same river system, 0.8 crabs/tow). Average size of adult crabs diverged between the North Bay population, which increased in size between 1996 and 2001, and the South Bay population, which decreased. The rapid establishment and spread of this species, its tolerance for a wide range of biotic and abiotic conditions, and its cyclical population dynamics pose challenges for control. Die in Flüssen und Ästuaren Zentralasiens beheimatete Chinesische Wollhandkrabbe hat im letzten Jahrhundert einige europäische Länder erobert und verursacht weit verbreitete Bedenken, aufgrund ihrer periodisch auftretenden extremen Häufigkeit und ihres Grabverhaltens, das Ufererosion verursacht. In der Bucht von San Francisco und ihren Zuflüssen befindet sich die erste und bisher einzige etablierte Population der Chinesischen Wollhandkrabbe in Nordamerika. Nach ihrer Entdeckung im Süden der Bucht von San Francisco 1992 hat sich die Wollhandkrabbe schnell über einige tausend Quadratkilometer in der Umgebung der Bucht ausgedehnt. Zwischen 1995 und 2001 untersuchten wir die Verbreitung, die Populationsdynamik und die Eigenschaften der Lebensweise der Chinesischen Wollhandkrabbe in der Bucht von San Francisco und ihren Zuflüssen. Die Dichte der Höhlen nahm in den Gezeitenbereichen der Zuflüsse in der südlichen Bucht von einem Mittelwert von 6 Höhlen/m2 1995 auf > 30 Höhlen/m2 1999 zu. Wollhandkrabben sind als junge Juvenile mit

*Corresponding author: Deborah A. Rudnick, Department of Environmental Science, Policy and Management, University of California, Berkeley, CA 94720-3112, Phone: +510-642-6315, Fax: +510-642-7428, E-mail: [email protected]

1439-1791/03/04/03-249 $ 15.00/0

250

Rudnick et al. gezeitengeprägten Bereichen der Zuflüsse der Bucht assoziiert, als ältere, wandernde Juvenile mit Süßwasserflüssen (≤ 250 Kilometer vom Zusammenfluss mit der Bucht) und als reproduktive Adulte mit dem offenen Wasser der Bucht, nachdem sie zwischen Spätherbst und zeitigem Frühjahr zur Reproduktion aus dem Süßwasser eingewandert sind. Die Populationsgröße erreichte 1998 mit 750.000 Krabben ihren Höhepunkt, die während der Herbstmigration in einem Zufluss der nördlichen Bucht gezählt wurden, sowie mit 2,5 Krabben/Schleppzug, die in den Brutarealen der nördlichen Bucht gesammelt wurden. Anschließend nahm die Häufigkeit sehr stark ab (2500 Krabben in dem gleichen Fluss, 0,8 Krabben/Schleppzug). Die durchschnittliche Größe der adulten Krabben unterschied sich zwischen der Population der nördlichen Bucht, deren Größe zwischen 1996 und 2001 zunahm, und der Population der südlichen Bucht, bei der sie abnahm. Die rasend schnelle Etablierung und Ausbreitung dieser Art, ihre Toleranz für einen weiten Bereich biotischer und abiotischer Faktoren und ihre zyklische Populationsdynamik stellen eine Herausforderung für ihre Kontrolle dar. Key words: introduced species – population dynamics – Crustacea – monitoring – life history – burrowing – bay – estuary

Introduction The Chinese mitten crab (Eriocheir sinensis H. Milne Edwards) was first discovered in San Francisco Bay in the winter of 1992. This non-native crustacean spread rapidly throughout the open waters of the Bay and its tributaries, giving this species the distinction of being one of the most abundant and widely distributed of the more than 200 introduced aquatic organisms found in the Bay (Cohen & Carlton 1998). The recent introduction of this species and its potential for widespread impacts makes E. sinensis a timely and important subject for the study of the processes of biological invasion in aquatic ecosystems. Eriocheir sinensis is catadromous; pelagic larvae are released into coastal waters and metamorphose into benthic juvenile crabs that migrate into brackish and fresh waters. The juvenile crabs spend between 1 and 5 years in freshwater streams and, at sexual maturity, migrate back to saline waters to reproduce. In Asia and Europe, adult, sexually mature crabs (i.e., females with developed secondary structures for carrying eggs; males with mature reproductive organs and/or accessory structures) begin to migrate downstream in late fall and winter. Female E. sinensis carry from 250,000 to 1,000,000 eggs attached to the underside of their abdomen. The crabs are believed to have one reproductive season and die shortly after reproduction (Panning 1939). One of the most often described features of E. sinensis is its propensity for burrowing. Juvenile mitten crabs create burrows in banks between the high and low tide lines in tidally influenced portions of streams, and occasionally in muddy banks of freshwater streams. In areas where crabs are particularly abundant, the burrows become tightly packed together and often are interconnected (Rudnick et al. 2000).

Basic Appl. Ecol. 4, 3 (2003)

High burrow densities of the Chinese mitten crab have been linked with bank and levee weakening, and even collapse in some areas into which this species has been introduced (Panning 1939, Dutton & Conroy 1998). The Chinese mitten crab is native to the rivers and estuaries of central Asia, from the west coast of North Korea to south of Shanghai, China (Panning 1939). The wild population of the crab has severely declined because of a combination of water pollution, dams that block migration to rearing areas, and commercial harvesting (Hymanson et al. 1999). This species has become established, at varying levels of abundance, throughout most of western, central, and northern Europe. In the 1930s, mitten crabs became a serious pest in Germany, causing damage to banks and levees through burrowing activities and to fishing operations through entanglement in nets and injury to netted fish (Panning 1939). The Chinese mitten crab has also been established in France, England, Denmark, Sweden, Finland, Luxembourg, Poland, Austria, Czechoslovakia, and most recently in Portugal and the Seville area of Spain (Hoestlandt 1944, Haahtela 1963, Anonymous 1972, Ingle 1986, Ja˙zd˙zewski & Konopacka 1993, Dhur & Massard 1995, Clark et al. 1998, Cabral & Costa 1999, Ferrero-Rodriguez pers. comm.). In the Baltic countries, however, the population densities have remained quite low, possibly because the low salinity and/or low temperature of the Baltic Sea inhibits successful reproduction of the Chinese mitten crab (Rassmussen 1987). Likely mechanisms of introduction of the crab to these countries include accidental importation through ballast water or spread to adjacent countries through streams or coastal waterways (Cohen & Carlton 1997). In North America, a few Chinese mitten crabs have been collected from the Mississippi River (Horwath

Eriocheir sinensis in San Francisco Bay

1989), Lake Erie (Nepszy & Leach 1973), and the Columbia River Basin (P. Heimowitz pers. comm.). None of these areas, however, currently supports a self-sustaining population of mitten crabs. In California, E. sinensis was first collected by commercial shrimp trawlers in the open waters of South San Francisco Bay in 1992, and the population quickly became established and abundant throughout San Francisco Bay and its tributaries. It is likely that mitten crabs arrived in the Bay, as well as in other parts of the continent, either through ballast water or as a result of release of live adult crabs imported for sale in Asian markets (Cohen & Carlton 1997). The objective of this article is to describe the population dynamics and ecology of the Chinese mitten crab in San Francisco Bay and its tributaries. We describe the invasion history, abundance, distribution, population structure, reproductive patterns, and habitat associations of the Chinese mitten crab throughout this region, and discuss the applications of this research to monitoring and management of this species.

251

Regional data We collected data on the spread of the mitten crab at a regional level on a continuous basis from 1992 through 1998. Regional distribution data covers all tributaries to San Francisco Bay, including the California Aqueduct and Delta-Mendota Canal that extend to central and southern California. In 1996, the California Department of Fish and Game (CDFG) began to receive numerous reports of collections and sightings of the crabs. These reports, coupled with results of trawls and distribution surveys described below, enabled us to produce maps to describe the distribution of the mitten crab in 1992, 1994, 1996, and 1998.

South San Francisco Bay data

Materials and Methods

South Bay data were based on: 1) juvenile mitten crabs that were monitored in tidal and freshwater portions of South Bay tributaries; and 2) adult mitten crabs that were collected from the open waters of the South Bay from weekly trawls conducted year-round by the Marine Science Institute (MSI), a nonprofit research and education organization based in the South Bay.

Our research efforts focused on three areas of northern California: 1) a regional study of the crab across San Francisco Bay and its tributaries; 2) studies in the tributaries and open waters of South San Francisco Bay; and 3) studies in San Pablo and Suisun Bays, Suisun Marsh, and the Sacramento-San Joaquin Delta (collectively referred to as the North Bay and Delta) (Fig. 1 and Table 1).

Juvenile mitten crab surveys In the freshwater tributaries of the South Bay, we conducted research on the population dynamics and habitat associations of juvenile mitten crabs. In 1996, 1999, and 2000, we monitored the distribution of juvenile crabs at 72 sites along 14 tributaries (Fig. 1, South Bay inset). Some of the tributaries are concretelined flood control channels that carry water only dur-

Table 1. Origin, methods, dates, and geographic coverage for Chinese mitten crab data Data collected by

Type of data collected

Methods

Years of data

Geographic coverage

University of California at Berkeley

Juvenile distribution and habitat association Juvenile burrow densities

Visual census, trapping 5 m transects, burrow excavation

1996, 1999, 2000 1995, 1996, 1999, 2000

South San Francisco Bay tributaries Tidally influenced portions of South Bay tributaries

Marine Science Institute

Adult population dynamics and morphology

Otter trawl

1995–2001

Open waters of South San Francisco Bay

California Department of Fish and Game

Juvenile burrow densities

5 m transects

1997, 1998

Adult population dynamics and morphology

Otter trawl

1996–2001

Suisun Marsh and tidally influenced portions of the Delta North Bay and Delta

University of California at Davis

Adult population dynamics and morphology

Otter trawl

1996–1999

Suisun Marsh

US Bureau of Reclamation, Tracy Fish Collection Facility

Adult population dynamics and morphology

Collections from fish salvage tanks

1996–2000

southern Delta

Basic Appl. Ecol. 4, 3 (2003)

252

Rudnick et al.

Fig. 1. Regions of study in the San Francisco Estuary. Abbrevations: CDFG = California Department of Fish and Game; UCB = University of California, Berkeley; UCD = University of California, Davis; MSI = Marine Science Institute.

Basic Appl. Ecol. 4, 3 (2003)

Eriocheir sinensis in San Francisco Bay

ing storm events; 20 of the 72 sites sampled were this type of concrete channel. A majority of tributaries monitored throughout the South Bay, however, are lined with natural materials, at least along most of their lengths, and confined by set-back levees that allow the river to occupy a restricted portion of its original floodplain. Sites were visited during late summer of each year and were chosen to represent a wide range of salinities, habitat types, and geographical locations throughout the South Bay watershed. The presence or absence of mitten crabs at each site was established by burrow excavation, setting crab traps, and visually inspecting the area for mitten crabs. Burrows were excavated by hand using trowels. Visual inspection entailed daytime searches for mitten crabs in channel margins, spending at least 30 minutes at a site before concluding that mitten crabs were likely absent. Habitat parameters associated with the presence or absence of mitten crabs were measured, including: average depth and temperature of water; bank substrate components (clay, silt, organic debris, or gravel); and bank vegetation composition and percent cover. Measurements were not made at sites that were concretelined flood control channels because the streams did not provide any natural habitat and were often devoid of water. In the tidal portions of these freshwater tributaries, we monitored abundance and distribution of mitten crab burrow colonies at four sites during the summer of 1995, 1996, 1999, and 2000 (Fig. 1, South Bay inset). Three sites used in 1995 and 1996 were no longer usable after they received substantial sediment deposition in 1997–98; these sites were substituted in 1999 and 2000 with alternative sites that were similar in salinity and distance from the Bay. The tidally influenced portions of South Bay tributaries range between 0 and 15 ppt in salinity and are up to 8 km in length. The muddy banks of these constrained channels provide ample habitat for juvenile mitten crabs to establish burrows. At each site, burrow density surveys were conducted during the early summer because juvenile mitten crabs are likely to be concentrated in brackish habitats during this time, prior to their upstream migration to freshwater habitats in the late summer and fall (Panning 1939). The same habitat parameters were measured at these sites as were collected for all other surveyed sites (see above). Burrows were used as a surrogate for estimating local crab abundance because they are easy to measure and provide evidence of crab presence regardless of whether the crab can be seen (Genoni 1991, Mouton & Felder 1996). Chinese mitten crab burrows are easily differentiated from other species’ burrows, such as crayfish, because the shape of the crab burrow entrances are elliptical, reflecting the flattened, oval

253

shape of the crab’s body. We randomly selected three transects at each site to measure mitten crab abundance; burrow density was calculated as the average of the three transects. Each transect was 5 m in length and ran parallel to the stream channel along the exposed vertical banks; these transects encompassed the distance between the low water line and the high tide mark along the bank. All burrows within each transect were counted to derive the number of burrows/m2 of bank area. We made estimates of burrow occupancy by conducting burrow excavations of 0.5 m2 bank area at each site in 1995, 1999, and 2000. Morphological measurements of crabs collected during surveys were recorded. Size was determined by measuring carapace width (CW) at the widest part of the carapace, just posterior to the fourth antero-lateral spine, and measured to the nearest millimeter with calipers. Sex was determined by the shape of the abdomen for all crabs ≥10 mm CW; in crabs 32 mm CW and males >39 mm CW. An annual North Bay and Delta E. sinensis catch per unit effort (CPUE) was calculated as the average number of crabs per tow from October through March of the following year. Samples were not collected upstream of San Pablo Bay in January 1997, and were not collected throughout the region in November

Basic Appl. Ecol. 4, 3 (2003)

and December 1999 and March 2001; CPUE was not adjusted for these missing data. Given the large numbers of crabs collected in other years during these months, the calculated CPUE for 1996–97, 1999–2000 and 2000–01 likely are underestimates. Adult mitten crab surveys: Suisun Marsh Crabs from UCD trawls were sexed and measured as described for the CDFG and South Bay surveys. Physical data collected at each station included temperature (°C) and salinity (ppt). Crabs were first collected in February 1996, but adult crabs migrating downstream were not collected until fall 1997. For this reason, only data from 1997 on were used for analyses. Adult mitten crab surveys: southern Delta The US Bureau of Reclamation Tracy Fish Collection Facility (TFCF) collects Chinese mitten crabs that enter the facility during their downstream migration. The TFCF is located just upstream of power turbines and is designed to collect fish and transfer them safely away from the turbines. The Bureau collected size and sex data on these crabs beginning in 1997; at least 500 crabs were measured during each migration season. In addition, time of collection was recorded by the facility in order to describe diel timing of migration patterns. The TFCF tracked the mitten crab population entering the facilities using 10-minute counts every two hours. Combined with counts of crabs collected at water intake points and specially installed traps in 1998 and 1999, daily estimates of the total number of crabs collected at these facilities were made.

Results Distribution and abundance Regional distribution The distribution of the Chinese mitten crab expanded dramatically over the course of this study. Crabs were restricted to the South Bay in 1992, were found throughout the open waters and immediate tributaries to the main body of the Bay in 1994, were distributed throughout the central and western Delta and South Bay tributaries by 1996, and had expanded their distribution to the Sacramento and San Joaquin Rivers and their tributaries by 1998 (Fig. 2). Juvenile distribution and abundance Between 1996 and 2000, juvenile crabs rapidly spread within the South Bay and its tributaries. In 1996, we observed mitten crabs several km upstream in South Bay tributaries; by 1999 and 2000, they occurred up to 50 km upstream from the Bay (Fig. 3).

Eriocheir sinensis in San Francisco Bay

Burrow densities of juvenile crabs in South Bay tributaries increased dramatically between 1996 and 1999 (Fig. 4). At the four South Bay sites examined for juvenile abundance in 1995, burrow densities ranged from 3 to 12 burrows/m2 (all sites: 6.3 ± 1.2, mean ± s.e.); densities were similar in 1996. In 1999, burrow densities ranged from 21/m2 to 39/m2 (all sites: 30.0 ± 1.6, mean ± s.e.). In 2000, the mean density of all surveyed sites (21.1 ± 2.1, mean ± s.e) declined significantly (p < 0.05, t test). Based on excavations, average occupancy rates of mitten crab burrows in South Bay sloughs decreased over time, from 80% in 1995, to 64% in 1999, and to 58% in 2000. The combined occupancy rates and burrow densities yielded estimated population densities of 5 crabs/m2 in 1995, 19/m2 in 1999, and 12/m2 in 2000 (Fig. 4). In the North Bay and Delta, densities of juvenile mitten crabs remained low throughout the sampling period (Fig. 5). Suisun Marsh densities in 1997 and 1998 ranged between 0.4 and 1.7 crabs/m2 in 1997, and 0.3 and 3.0 crabs/m2 in 1998 (Fig. 5). In the Delta, densities ranged between 0 and 0.4 crabs/m2 in 1997, and 0 and 1.0 crabs/m2 in 1998.

255

Adult distribution and abundance Between 1995 and spring of 1999, abundance of adult mitten crabs in South San Francisco Bay was relatively constant. The highest catch per unit effort (CPUE), nearly double that of any previous year, was in the 1999–2000 reproductive season (Fig. 6). Adult crabs were first collected in the North Bay and Delta by CDFG trawls in fall 1996; abundance of crabs there increased dramatically in the 1998–99 reproductive season, and then declined in 1999–2000 (Fig. 6). Collections of adult, downstream-migrating crabs at the TFCF in the southern Delta reflected the same trends in abundance as seen in North Bay and Delta trawl data.

Fig. 3. Changes of distribution of juvenile crabs in the South San Francisco Bay catchment, 1996–2000; see Fig. 1 for the location of this map.

Fig. 2. Changes in regional distribution, indicated in bold black, of the Chinese mitten crab in the San Francisco Bay catchment, 1992–1998.

Fig. 4. Average burrow densities of juvenile crabs in South Bay tributaries: error bars are ± 1 s.e; 4 sites were sampled each year. Crab density estimates were generated by multiplying burrow occupancy rates by burrow densities. No occupancy rate data were collected in 1996.

Basic Appl. Ecol. 4, 3 (2003)

256

Rudnick et al.

Fewer than 100 crabs were collected at the facility in the fall of 1996. However, approximately 20,000 crabs were collected at TFCF during the fall of 1997, and an estimated 750,000 crabs were collected at TFCF in the fall of 1998. In 1999, collections at the facility declined dramatically and remained low in 2000 (Fig.6). Population structure

Fig. 5. Densities per m2 of juvenile crabs in Suisun Marsh and in three regions of the Delta. Error bars are ± 1 s.e; numbers above the bars are number of sites sampled in each region in each year.

Fig. 6. Number of adult crabs caught in three areas of the San Francisco Bay catchment; data is grouped from fall of one year to summer of the following year to incorporate the entire reproductive season into a sampling period.

At the four tidally influenced, burrow-monitoring sites in the South Bay, juvenile mitten crabs ranged in size from 5 to 35 mm CW (mean ± s.e. = 18 ± 0.6 mm, n = 230). Crabs > 35 mm CW and up to 80 mm CW were found at freshwater survey sites. Adult, sexually mature crabs collected in the open waters of the South Bay ranged in size from 30 mm CW to 80 mm CW; adults collected from North Bay trawls ranged in size from 33 mm to 84 mm CW; downstreammigrating adults collected at the TFCF ranged in size from 40 mm to 95 mm CW. Size did not differ significantly by sex for this population, nor did size vary significantly within the reproductive season for either sex. Over the first five years of collections from South San Francisco Bay, there was a continual decline in the yearly average size of adult crabs; in contrast, the mean size of all adult crabs collected in the North Bay and southern Delta increased from 1997 to1999 and showed little change in 2000–01 (Fig. 7). North Bay and southern Delta collections were significantly larger than the average sizes of crabs collected in the South Bay between 1998 and 2001 (ANOVA, Tukey HSD p < 0.001) (Fig. 7). The sex ratio of adult mitten crabs collected from both the South Bay and North Bay and Delta exhibited a consistent pattern of a greater proportion of males in the fall and early winter giving way to a higher proportion of females in the spring (Fig. 8). During nearly all years of the study period, the sex ratio of collected crabs was slightly biased towards males: in the South Bay, 55% of collected crabs were male; in North Bay and Delta trawls, 58% of all adults collected were males; and adult males collected from Suisun Marsh slightly outnumbered females (52%). Reproductive patterns

Fig. 7. Average size (mean ± 1 s.e.) of Chinese mitten crab adults during the reproductive season in three areas of the San Francisco Bay catchment, 1995–2001: North Bay and Delta and South Delta crabs (b) are significantly larger than South Bay crabs (a) in years 1998–2001 (Tukey HSD, p < 0.001).

Basic Appl. Ecol. 4, 3 (2003)

In the southern Delta, adult mitten crabs migrating downstream to their reproductive grounds were collected nearly exclusively in the fall months. Mitten crabs arrived in low numbers at TFCF in late August or early September during all years of monitoring, with daily catches peaking between mid-September and early October (Fig. 9). Downstream crab migration through TFCF showed a strong diel pattern. Mitten crabs entered the holding tanks at the highest rates

Eriocheir sinensis in San Francisco Bay

during the late evening and early morning hours; entries declined during the day (Fig. 10). Adult mitten crabs in the North Bay and Delta moved from shallower, fresher waters in the fall to deeper, more saline waters in winter and early spring. During the late summer through the early reproductive season (October-November), adults were most common in the lower rivers and in the shallow waters of Suisun Bay. Over December and January, there was a continued migration downstream to San Pablo Bay (Fig. 11). Adult mitten crabs were most commonly collected during the winter months in the deepest waters of San Pablo Bay. Adult males and non-ovigerous females continued to be collected upstream of San Pablo Bay after January. Most ovigerous females, however, were collected in San Pablo Bay; only 9% were collected upstream of San Pablo Bay and none were collected upstream of Suisun Marsh. From April–August, a few non-ovigerous females were collected in San Pablo Bay. We assume that these females were post-reproduction be-

257

cause many appeared to be sluggish, missing appendages and covered in epizooids, indicating that they had been in salt water for several months. Ovigerous females were collected from October to June, with the highest percentage of ovigerous females collected in the late fall and early winter of each reproductive season (Fig. 8). Over one-half of all females collected between November and April from both the North Bay and Delta and the South Bay were ovigerous. Habitat associations We found mitten crabs throughout three different habitats in San Francisco Bay and its tributaries: tidal channels (tributaries influenced by tidal inflow); freshwater creeks (tributaries upstream of tidal influence); and the open waters of the South, Central, San Pablo, and Suisun Bays. Salinity differed among the three habitats, ranging from 1–15 ppt in tidal sloughs,